Dwindling petroleum reserves and increasing prices have placed an increased emphasis on the use of synthesis gas in place of oil as a starting material for various chemicals such as methanol, formaldehyde and ethylene glycol. The advantage of synthesis gas is that it can be produced from raw materials other than petroleum such as natural gas or coal, and potentially from oil shale and tar sands.
An example of an industrial process for the production of ethylene glycol utilizing synthesis gas as a starting material is the reaction of formaldehyde with carbon monoxide and water at high pressures (over 300 atmospheres) in the presence of an acid catalyst to produce hydroxyacetic (glycolic) acid which is then reacted with methanol to give the methyl ester; the latter is then converted to the glycol by catalytic hydrogenation. See U.S. Pat. Nos. 2,316,564 issued Apr. 13, 1943 to Cockerill; 2,153,064 issued Apr. 4, 1939 to Larson; and 2,152,852; 2,385,448 and 2,331,094 issued Apr. 4, 1939, June 9, 1942 and Oct. 5, 1943, respectively, to Loder.
Another proposed process utilizing synthesis gas for the production of ethylene glycol is the reaction of methanol and carbon monoxide using a rhodium-catalyzed, high pressure process; see U.S. Pat. Nos. 4,115,428 issued to Vidal et al and 4,115,433 issued to Cosby et al on Sept. 19, 1978.
With respect to the type of process for the production of ethylene glycol disclosed and claimed herein, it should be noted that the oxidative dimerization or dehydrodimerization of a large variety of organic compounds by peroxides is very old art that was pioneered by the preeminent free radical theoretician M. S. Kharasch and his students. These studies became the foundations of much subsequent free radical chemistry. Karasch et al in JACS 65, 15, 1943 shows the dehydrodimerization of acetic acid to succinic acid with acetyl peroxide in a 50 mole percent utilization selectivity based on acetyl peroxide, utilization selectivity being defined as the moles of dehydrodimer product made divided by the moles of peroxide converted. Isobutyric acid produced tetramethyl succinic acid in a 42.4 mole percent utilization selectivity. Kharasch et al in J. Org. Chem. 10, 386, 1945 show the ester methyl chloroacetate being dimerized to dimethyl dichloro succinate by acetyl peroxide in a 41 percent utilization selectivity. Kharasch et al in J. Org. Chem. 10, 401, 1945 show the dimerization of cumene and ethyl benezene with acetyl peroxide in 61.9 mole percent and 32.1 mole percent respectively to their dehydrodimers. Wiles et al in I, E & C, August 1949, page 1682, tell of the efficacy of di-t-butyl peroxide and 2,2 bis (t-butyl peroxy) butane for the dimerization of cumene. The benzoate ester of benzyl alcohol was dimerized to the dibenzoate ester of the corresponding glycol, diphenylene glycol, with di-t-butyl peroxide by Rust et al JACS 70, 3258 (1948).
The literature is replete with many other examples that produced the dehydrodimers at very low concentrations at utilization selectivities of generally from 20-50 mole percent based on the peroxide consumed. These selectivities are generally too low for the process to be considered for commercial development.
In connection with ethylene glycol, two teachings involving peroxide-induced reactions should be mentioned.
The first is described by Schwetlick et al, Angew. Chem. 72, 1960, No. 21, pages 779 and 780, and involves heating a mixture of di-tertiary butyl peroxide and methanol in a molar ratio of 1:20 in an autoclave and/or under reflux for a period of 10 hours at 140.degree. C. A 26 percent yield of ethylene glycol is reported, with the statement being made that an increase in the alcohol excess raises the yields.
The second and more important of such other reaction paths to ethylene glycol, in terms of its relevance to the present invention, is described by Oyama in J. Org. Chem. 30, July, 1965, pages 2429-2432. In particular, Oyama shows the reaction of 9 moles of methanol, 1.8 moles of 15 percent aqueous formaldehyde and 0.45 moles of t-butyl peroxide (di-tertiary butyl peroxide) at 140.degree. C. for 12 hours to give 0.21 moles of ethylene glycol (Table I at the top of the right hand column on page 2430), with the statement being made immediately below Table I: "The yield of ethylene glycol in the reaction of formaldehyde with methanol is higher than that of t-butyl peroxide induced dimerization of methanol. This fact suggests that hydroxymethyl radical (D) adds to formaldehyde." Oyama describes in greater detail how this reaction was run and the products obtained, and contrasts it with the dehydrodimerization of methanol in the presence of t-butyl peroxide and the absence of formaldehyde, in the "Experimental" section beginning at page 2431 (particularly the sections headed "Reaction of Methanol with Formaldehyde" and "Dimerization of Methanol" on page 2432).
The yields of ethylene glycol obtained by Oyama are fairly low. Oyama's only run with methanol--that involving the above-described reaction of methanol, aqueous formaldehyde and t-butyl peroxide at 140.degree. C. for 12 hours--gave only 1.86 weight percent of ethylene glycol.
The above-described reaction can be made to produce higher yields of ethylene glycol by substantially decreasing the amount of organic peroxide employed, relative to the amounts of formaldehyde and methanol present, from that employed by Oyama. Moreover, increasing the amount of methanol and decreasing the amount of water, relative to the other components of the reaction mixture, in contrast to the amounts employed by Oyama, also appear to contribute to the production of higher yields of ethylene glycol. Thus, for example, heating a mixture of 78.5 weight percent of methanol, 1.5 weight percent of di-tertiary butyl peroxide, 6.9 weight percent of formaldehyde and 13.1 weight percent of water at 155.degree. C. for 2 hours gave a yield of 4.5 weight percent of ethylene glycol in the product mixture. This is equivalent to a yield of about 7.1 moles of ethylene glycol per mole of di-tertiary-butyl peroxide employed. (Oyama obtained 0.466 mole of ethylene glycol per mole of di-tertiary-butyl peroxide in his reaction). This improvement is more fully disclosed in the copending parent of this application, U.S. Ser. No. 183,537, filed Sept. 2, 1980.